Update README.md
Browse files
README.md
CHANGED
|
@@ -17,6 +17,12 @@ Previous methods largely rely on in-context approaches to generate instructions,
|
|
| 17 |
Inspired by this, we took this approach further by grounding the generations in user-provided context passages.
|
| 18 |
Further, the model is trained to generate questions involving complex scenarios that require detailed reasoning, allowing for models trained on the generated data to reason step-by-step.
|
| 19 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 20 |
An example notebook is provided [here](https://huggingface.co/NousResearch/Genstruct-7B/blob/main/notebook.ipynb), which details how to load and sample from the model.
|
| 21 |
|
| 22 |
Alternatively, here's a minimal example:
|
|
|
|
| 17 |
Inspired by this, we took this approach further by grounding the generations in user-provided context passages.
|
| 18 |
Further, the model is trained to generate questions involving complex scenarios that require detailed reasoning, allowing for models trained on the generated data to reason step-by-step.
|
| 19 |
|
| 20 |
+
| | ChatGPT | Few-shot prompting | RAG | Ada-Instruct | **Genstruct** |
|
| 21 |
+
|:-------------------:|:--------:|:------------------:|:---:|:------------:|:---------:|
|
| 22 |
+
| Open models | ❌ | ☑️ | ☑️ | ✅ | ✅ |
|
| 23 |
+
| Grounded generation | ❌ | ❌ | ✅ | ❌ | ✅ |
|
| 24 |
+
| Complex reasoning | ✅ | ☑️ | ❌ | ☑️ | ✅ |
|
| 25 |
+
|
| 26 |
An example notebook is provided [here](https://huggingface.co/NousResearch/Genstruct-7B/blob/main/notebook.ipynb), which details how to load and sample from the model.
|
| 27 |
|
| 28 |
Alternatively, here's a minimal example:
|